[Mike Monett]

74AC74 oscillator? I made an oscillator from a flipflop once, but it
took a lot of parts and broke some rules.

Here's an approach that breaks no rules. It uses two caps, two resistors,
one diode, two ic's, and provides a symmetrical duty cycle with adjustable
dead time.

Tie the Clk and D of a 74AC74 to ground. Connect the main timing cap from
the Set to VCC. Add a timing resistor R1 from Set to Q. Add a diode in
parallel with R1 and the anode to Q.

Qbar and Reset are the same without the diode. The R2C2 time constant can
be 5% of the main R1C1 to adjust the dead time.

Feed the Q output to the Clk of the remaining flop connected as a divide by
two. Feed QBar to a 74AC00, which gates the output of the divide by two
flop. When the gate output goes low, turn on that side of the bridge. A
74AC32 could be used to invert the drive polarity.

I can do it (and probably will) with a sot-23 schmitt/rc oscillator,
an sc-70 flipflop, and an so-8 dual mosfet gate driver (a little one,
rated 4 amps per side), total 5 parts and about 1 cm square, plus a
couple of bypasses somewhere. Oh, there's +5 available for other
reasons.

If I got brave, I'd make the osc with just the schmitt gate and a
resistor, no cap.

And some more pcb flux cleaning issues?

Maxim makes an oscillator-flipflop-driver, but the outputs are
open-drain, and that won't work with my transformer.

I'm trying to get a full H-bridge working using Terry's approach and a pair
of IRF3706's and IRF7410's. Unfortunately the SPICE model uses subckts and
I can't get MC8 to turn on the P-channel MOSFET. But that would reduce the
parts count significantly.

What I want is to derive some regulated -5 for opamps and something
like +20 to +30 to bias the photodiode. The pin photodiodes get faster
as the backbias increases, partly because the capacitance goes down
but mainly because the increased field sweeps carriers out faster.

So, 10 volt diff drive into a "1" winding. The other "1" winding gets
bridge rectified to -8, then lin reg'd to -5. The two "2" windings are
seriesed and diode rectified to +40ish, then zenered down to +28 for
the pin bias, something like that.

The other posts also look interesting.

I have quantities of the IDSN transformer, so I'd like to use it here.
It's cute, about a cm cube, surface mount, under a dollar each. What I
don't have is a lot of space.

Seems to be a constant problem with most people

John

Regards,

Mike Monett

 

[John Larkin]

And some more pcb flux cleaning issues?

If I get nA (and sometimes pA) leakages, almost none of my boards
work, so clean is a given. But I don't know how predictable the
frequency would be if it depended on device capacitance for the "c".
That's too much risk to zap one 0603 cap, I guess.

I'm trying to get a full H-bridge working using Terry's approach and a pair
of IRF3706's and IRF7410's. Unfortunately the SPICE model uses subckts and
I can't get MC8 to turn on the P-channel MOSFET. But that would reduce the
parts count significantly.

Buying the SO-8 dual mosfet driver reduces parts count more! Circuit
"design" is, more and more, a matter of just buying the right chip.

I did find my 1310 nm single-mode laser, $36 in small qtys. These guys
seem very nice...

http://www.appointech.com/fomodules.html


John

 

[Clifford Heath]

its piss easy, I shouldnt need to draw a picture.

Yup, got it, thanks.

A circuit like yours is shown here (p19) but with two NFETs instead:
<http://schmidt-walter.fbe.fh-darmstadt.de/snt/snt_eng/snteng4.pdf>.
Without really thinking it through, it looks like you'd need to be
careful with the gate drive of the upper FET though.

In your design, an input voltage spike could kill the PFET.
Do you clamp the PFET gate to V+ with a zener?

Clifford Heath.

 

[Terry Given]

Yup, got it, thanks.

A circuit like yours is shown here (p19) but with two NFETs instead:
<http://schmidt-walter.fbe.fh-darmstadt.de/snt/snt_eng/snteng4.pdf>.
exactly

Without really thinking it through, it looks like you'd need to be
careful with the gate drive of the upper FET though.

not really. as long as the bottom FET is well driven, it doesnt matter
too much what the top FET does. lousy gatedrive simply limits maximum
duty cycle, which you can compensate for with turns ratio.

In your design, an input voltage spike could kill the PFET.
Do you clamp the PFET gate to V+ with a zener?

Clifford Heath.

the P-FET gate is clamped to (+Vdc + Vf) max by the N-FETs commutating
diode, and -Vf min by the N-FETs D-S diode.

Yes if +Vdc is too high you can break the FET, but if you are going to
use +12V/+15V then the FET rated Vgmax will be +/-30V anyway, and as JL
showed you really have to work at it to overvoltage a FET gate. The +Vdc
supply rail ought to have some reasonable amount of capacitance, which
bounds the spike energy/clamps the spike voltage - if you have 100uF on
a 15Vdc bus and dont want it to go above 30V, then dE = 50uF*(30V^2 -
15V^2) = .033J = 1500pF charged to 6,700V, so you can see a decent bus
cap alone can deal with an ESD blast.

Cheers
Terry

 

[Terry Given]

74AC74 oscillator? I made an oscillator from a flipflop once, but it
took a lot of parts and broke some rules.

I can do it (and probably will) with a sot-23 schmitt/rc oscillator,
an sc-70 flipflop, and an so-8 dual mosfet gate driver (a little one,
rated 4 amps per side), total 5 parts and about 1 cm square, plus a
couple of bypasses somewhere. Oh, there's +5 available for other
reasons.

I'm too stingy to buy those cute little single-gate things, but you use
plenty of them. I like it.

If I got brave, I'd make the osc with just the schmitt gate and a
resistor, no cap.

that'd be a no-no for me. I sit these things right on top of nasty,
nasty stuff. I tend to use pullups no bigger than 10k, and capacitively
load every input as much as it will tolerate. Along with solid 0V
planes, etc.

Maxim makes an oscillator-flipflop-driver, but the outputs are
open-drain, and that won't work with my transformer.

What I want is to derive some regulated -5 for opamps and something
like +20 to +30 to bias the photodiode. The pin photodiodes get faster
as the backbias increases, partly because the capacitance goes down
but mainly because the increased field sweeps carriers out faster.

So, 10 volt diff drive into a "1" winding. The other "1" winding gets
bridge rectified to -8, then lin reg'd to -5. The two "2" windings are
seriesed and diode rectified to +40ish, then zenered down to +28 for
the pin bias, something like that.

I have quantities of the IDSN transformer, so I'd like to use it here.
It's cute, about a cm cube, surface mount, under a dollar each. What I
don't have is a lot of space.

John

Hi John,

do you have a link to the transformer datasheet?

Cheers
Terry

 

[John Larkin]

I'm too stingy to buy those cute little single-gate things, but you use
plenty of them. I like it.


that'd be a no-no for me. I sit these things right on top of nasty,
nasty stuff. I tend to use pullups no bigger than 10k, and capacitively
load every input as much as it will tolerate. Along with solid 0V
planes, etc.



Hi John,

do you have a link to the transformer datasheet?

Cheers
Terry

We're using the SMJ-140, but they have lots of other ratios...

http://www.alfamag.com/PDF/TelecomsNetwork/ISDN/S-Interface Single/So-J_Trafo_Jul-03.pdf


These are made in India, so take about 4 weeks to get. Cost is about
$1 at 1K pieces. They are very nice, no quality problems so far. They
have a nice high inductance, low dcr, low leakage inductance, so are
good for tiny inverters in the low watts range. And there are other
sources for the same parts, just in case. See pic in abse.

I'm thinking the L is sorta high for most flyback apps.

I never though that ISDN made much sense, but the transformers are
great.


John

 

[Harry Dellamano]

We're using the SMJ-140, but they have lots of other ratios...

http://www.alfamag.com/PDF/TelecomsNetwork/ISDN/S-Interface Single/So-J_Trafo_Jul-03.pdf


These are made in India, so take about 4 weeks to get. Cost is about
$1 at 1K pieces. They are very nice, no quality problems so far. They
have a nice high inductance, low dcr, low leakage inductance, so are
good for tiny inverters in the low watts range. And there are other
sources for the same parts, just in case. See pic in abse.

I'm thinking the L is sorta high for most flyback apps.

I never though that ISDN made much sense, but the transformers are
great.


John

John,
I think that you need some power supply help, just how much power is
required on the +30V and -5V rails? That ISDN transformer is large enough
for maybe 30 watts! The primary inductance is so large that it cannot be
used as an inductor in a switcher. Idc is 5mA and 2R0 windings, so what is
that....150mW? 30mH and 100pF is self resonant at 91KHz so higher frequency
is out of the question. Winding AC resistance is probably >20R0. Best bet,
short out the primary windings and use the leakage inductance in a boost
switcher.
Let's have those power requirements and redesign with low cost inductors.
I bet Terry must of gasp when he saw that transformer spec. When I see 30mH,
I am thinking of >100 volts. Lots of turns for wasted inductance.
Harry

 

[Terry Given]

We're using the SMJ-140, but they have lots of other ratios...

http://www.alfamag.com/PDF/TelecomsNetwork/ISDN/S-Interface Single/So-J_Trafo_Jul-03.pdf


These are made in India, so take about 4 weeks to get. Cost is about
$1 at 1K pieces. They are very nice, no quality problems so far. They
have a nice high inductance, low dcr, low leakage inductance, so are
good for tiny inverters in the low watts range. And there are other
sources for the same parts, just in case. See pic in abse.

I'm thinking the L is sorta high for most flyback apps.

I never though that ISDN made much sense, but the transformers are
great.

John

For the design I am currently working on, I have gone and re-arranged
the relevant DCM flyback equations, and now choose Bmax and solve for
Fsmps given Psmsp. Then I crank up Psmps to include necessary
tolerances, and voila.

but 30mH is a bit high to do much of anything with for a flyback, unless
the voltage is very high and the frequency is very low.

I'd use that puppy as a crude dc-dc, but slap a PCMC loop around it
(which is essentially free) to ensure startup is nicely controlled. Or
use naff all output C, such that dTj is acceptable for your switching
transistor(s).

Cheers
Terry

 

[Didi]

John,

you may want to look at the MC34151 (MC33151) dual MOSFET driver.
One of them can oscillate while the other could drive your coil.
There was an application note 15 years ago on something similar
(just voltage doubler, you will need only the oscillator of it), if you
cannot locate it let me know and I'll dig my implementation of it
for you. I guess with this biasing voltage for a PIN diode you are
doing something which I might see as competitor stuff, but then
my help is only minute and aged enough .

Dimiter

 

[Ken Smith]

I want to drive both ends of one untapped winding, so what would work
would be an oscillator, a flipflop, and a pair of rail-rail drivers,
like maybe a full-bridge thingie or a dual mosfet gate driver. Current
will be low, 100 mA maybe.

Just a thought:

If the transformer is made into a series tuned circuit it could be driven
single ended and still make as high of a voltage. This assumes you get a
"Q" of at leat 2.

If the drive comes from a PWM controller chip, you could get regulation by
moving the duty cycle away from 50-50. There are quite a few controller
chips that have a 50-50 upper limit on the duty cycle. You may be able to
make something that regulates this way.

 

[Terry Given]

John,
I think that you need some power supply help, just how much power is
required on the +30V and -5V rails? That ISDN transformer is large enough
for maybe 30 watts! The primary inductance is so large that it cannot be
used as an inductor in a switcher. Idc is 5mA and 2R0 windings, so what is
that....150mW? 30mH and 100pF is self resonant at 91KHz so higher frequency
is out of the question. Winding AC resistance is probably >20R0. Best bet,
short out the primary windings and use the leakage inductance in a boost
switcher.
Let's have those power requirements and redesign with low cost inductors.
I bet Terry must of gasp when he saw that transformer spec. When I see 30mH,
I am thinking of >100 volts. Lots of turns for wasted inductance.
Harry

He didnt gasp, he banged a few numkbers into an equation and went "Hmm,
100Hz for 1W flyback at 5V" and immediately abandoned trying to use ot
for a flyback

Alas the voltage rating isnt quite enough for me. I'll stick to kool-mu
toroids, so I can get my 6kV isolation.

Cheers
Terry

 

[John Larkin]

John,
I think that you need some power supply help,

Yes, I recall asking for help at the start of this thread. I need an
oscillator, a flipflop, and dual 10-volt totem-pole driver, and I was
wondering if I missed something cute out there. Apparently not.

just how much power is
required on the +30V and -5V rails?

+30, maybe 1 mA; -5, maybe as much as 70 mA, certainly no more.

That ISDN transformer is large enough
for maybe 30 watts!

Maybe if you water cooled it.

The primary inductance is so large that it cannot be
used as an inductor in a switcher.

Funny, somebody else already made that comment. Me.

Idc is 5mA and 2R0 windings, so what is
that....150mW? 30mH and 100pF is self resonant at 91KHz so higher frequency
is out of the question.

If you're hung up on "flyback", yes. But there *are* other topologies.

We already use these transformers in two products...

http://www.highlandtechnology.com/DSS/V470DS.html

http://www.highlandtechnology.com/DSS/V450DS.html

with a couple more in the works. They work great. The classic
push-pull square-wave inverter is quiet, efficient, and has decent
regulation and an inherently low output impedance, even run open-loop,
as these are.

Let's have those power requirements and redesign with low cost inductors.

$1 is low cost here. And we have reels of these xfmrs in stock.

I bet Terry must of gasp when he saw that transformer spec. When I see 30mH,
I am thinking of >100 volts. Lots of turns for wasted inductance.

Whoops, the timer just dinged. My custard bread pudding is done.

John

 

[John Larkin]

He didnt gasp, he banged a few numkbers into an equation and went "Hmm,
100Hz for 1W flyback at 5V" and immediately abandoned trying to use ot
for a flyback

What is this thing people have for flybacks? Nasty, noisy, badly
regulated things!

John

 

[Didi]

What is this thing people have for flybacks? Nasty, noisy, badly

regulated things!

John

Not really, if you know how to make them. I have no problem powering
14 bit 10 MSPS ADCs with millivolts fullscale range at the input, and
powering the HPGe detector preamp off my convertors, all flyback (well,
I use step-down for some of the voltages, where applicable). The
resolution
(and the throughput, btw) of my systems beats that of any heavy
NIM monsters or whatever with linear supplies. In a device I have
all the +/-15, +/-24 etc. lower voltages one cannot do without, those

5V

are all flyback convertors.
Obviously I am not recommending you to do it, if you have not done it
yet. I am only setting the record straight, flyback convertors can be
used and have been used in very low noise circuits.

Dimiter

 

[Terry Given]

What is this thing people have for flybacks? Nasty, noisy, badly
regulated things!

John

I like them for several reasons. With a diagonal half-bridge I get
fairly good open-loop regulation from no load to full load, with fairly
good efficiency, and can do that across an evil isolation barrier (think
3300V IGBTs etc).

Secondly, if I use a sufficiently high voltage secondary diode, I can
stand off a worst-case fault (full dc bus volts across gate drive)
without snotting the primary circuitry.

and of course I can operate continuously into a dead short, no worries.
easy to detect also.

I'm designing a 250W unity power factor universal input flyback at the
moment, along with a 20W flyback that runs from 100Vdc to 1400Vdc.

yeah they can be noisy, but only if you dont know what you are doing.
not so good for low voltage high power, unless you crank up the
frequency enough to use ceramic caps.

Cheers
Terry

 

[[email protected]]

Hi,

I have a nice 4-winding ISDN transformer I want to use to generate
some odd supply voltages from a +10 basic supply. I need a biggish
voltage to bias a PIN photodiode, and some -5 to run some analog
stuff. The transformer is 1:1:2:2, so there are possibilities.

I want to drive both ends of one untapped winding, so what would work
would be an oscillator, a flipflop, and a pair of rail-rail drivers,
like maybe a full-bridge thingie or a dual mosfet gate driver. Current
will be low, 100 mA maybe.

So, anybody know of a single, tiny chip that might do all this?
(Flyback wouldn't work well with this xfmr.)

John

SG3525

 

[John Larkin]

I like them for several reasons. With a diagonal half-bridge I get
fairly good open-loop regulation from no load to full load, with fairly
good efficiency, and can do that across an evil isolation barrier (think
3300V IGBTs etc).

Half-bridge on the drive side? That's not a flyback.

John

 

[John Larkin]

Not really, if you know how to make them. I have no problem powering
14 bit 10 MSPS ADCs with millivolts fullscale range at the input, and
powering the HPGe detector preamp off my convertors, all flyback (well,
I use step-down for some of the voltages, where applicable). The
resolution
(and the throughput, btw) of my systems beats that of any heavy
NIM monsters or whatever with linear supplies. In a device I have
all the +/-15, +/-24 etc. lower voltages one cannot do without, those
are all flyback convertors.
Obviously I am not recommending you to do it, if you have not done it
yet.

Of course I've done it, and posted HV flyback designs to this NG
before your time. The push-pull square-wave converter has a nice
smooth input current flow, close to 100% output conduction (hardly
needs filtering), is quiet, and has load regulation so and in my app
needs no feedback to the driver. In a multiple-channel config (like
the pic I posted to abse) you only need a single simple complememtary
square-wave generator as the common gate drive for all channels, and
each channel only needs a pair of 2N7002's (6 cents total) to drive a
ct'd primary... and no channel-channel sync problems! Adding series
gate resistors softens up the edges - nice trapezoids - and greatly
reduces spikes, somethning that's not as nicely doable in a flyback.

Really, lots of people seem to have an unreasonable affection for
flybacks.


John

 

[John Larkin]

SG3525

Yikes, I'd clean forgotten about that one... haven't used that in,
well, a long time. I can probably tolerate the small voltage loss in
the bipolar output stage. Thanks.

John

 

[John Popelish]

Half-bridge on the drive side? That's not a flyback.

What could you do with these?
http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=7598144809